Combined neural network and particle filter state estimation with application to a run-of-mine ore mill

A run-of-mine (ROM) ore milling circuit poses many difficulties in terms of measuring process variables and determining accurate models. Control of the ROM circuit is therefore not a trivial task to achieve. An example of a ROM circuit model with reduced complexity that works well for control purposes is discussed. The mill model is discussed in detail, as this model is used for state estimation. A neural network is trained with three disturbance parameters and used to estimate the internal states of the mill, and the results are compared with those of particle filter implementation. A novel combined neural network and particle filter state estimator is presented. The estimation performance of the neural network is promising when the disturbance magnitude used is smaller than that used to train the network.